MTGO’s ‘Treasure Chests’ and competitive online play

Weighing in on another controversy here.

There’s a lot of unease or even doom and gloom about the MTGO prize changes.

I’m not someone who jumps on board every MTGO change. At the time I was quite hostile to the introduction of Play Points, although MTGO later added enough PP sinks to allay my fears.

That said, I think the Treasure Chest system is actually fairly positive overall – with caveats and room for improvement.


I’m not talking about the redemption changes here, other than a three paragraph aside (skip ahead if you want to avoid a rant). While I disagree with the prevailing opinions on the Treasure Chest system, the redemption changes are a steaming pile of shit that will enrich speculative traders at the expense of Standard players.

In the short term the redemption changes will also seriously hurt the high-volume low-margin bot chain traders that (in the absence of a good in-client MTGO trade system) serve as the lubricant that makes MTGO so smooth. Dealers like MTGOtraders, Goatbots, Clan Team and Cardhoarder have the largest collections on MTGO, and so changes that reduce consumer confidence in the MTGO platform (and hence reduce the secondary market value of Event Tickets in USD) hurt them more than anyone else. If tickets fall from USD 0.98 to USD 0.92, this might cost my collection USD 200 but might cost Goatbots USD 10000 or more.

Many people think ‘I never redeem so this doesn’t affect me’, but people that played MTGO when Zendikar was the newest set know how much restrictions on redemption fuck over players that have never redeemed a set and have no plans to ever do so. Those players saw their collections lose a third of their value during the redemption freeze.

But that’s enough ranting for now.



On MTGO, any prize that is available from Constructed events can and will flood the market unless it is desirable in extremely large quantities. As an example, Theros boosters spent much of their second year in Standard valued at around 80 cents.

It wasn’t that people didn’t want Theros packs. Drafters destroyed them all the time, even after the focus moved on to KTK drafts. Theros still had some cards that were in considerable demand and after FRF launched, it had a higher EV than KTK packs.

But there had just been too many of them awarded for drafters to absorb them all.

The most common reaction to Treasure Chests is to assume the same will happen. I dispute this. The very high rarity of the most desirable cards will prevent this happening, even with a large number of Chests being paid out as prizes.


So far the only information we have on Treasure Chest contents is the announcement, plus a video of Lee Sharpe opening ten of them.

The announcement states the following:

  • Specific rares (e.g. Snapcaster) are twice as common as specific mythics (e.g. Godsire) in the ‘Modern legal rare slot’ found in some treasure chests.
  • Curated cards are in different rarity tiers and there is no information on these tiers.
  • You get an average of 1.25 of the ‘good items’ (Modern rares/mythics, bundles of PP, curated cards) per chest, with a minimum of 1, and an average 1.75 of the ‘chaff’ (Standard commons/uncommons).

In the ten opened chests, Lee gets 2 curated cards, 6 modern legal rares/mythics, and 50 PP. While this is not enough information to determine the rarity of each of those prizes, I will make the assumption that this prize pool was fairly typical, and so will assume the following distribution:

  • 50% of the ‘good items’ are Modern rares/mythics
  • 35% are 10 Play Point bundles
  • 15% are curated cards.

This distribution would make the probability of opening a specific Modern rare (for example, 8th Edition Ensnaring Bridge) at around 1 per 5000 Treasure Chests. 1 per 10000 for a specific mythic (for example, Magic 2012 Time Reversal).

There is not enough information to determine much about the rarity of individual curated cards. However, given Wizards’ past practice with PZ1 (the most similar set conceptually), it’s reasonable to estimate that the most desireable curated cards will be 5 times as rare as the non-chase ones. (For example, PZ1 True-Name Nemesis, a mythic, is about 5 times as rare as PZ1 Arcane Sanctum, an uncommon).

If this is accurate, it’s plausible that Rishadan Port is 1 per 15000 to 1 per 20000 Treasure Chests.

This won’t trash the card’s value overnight like the Lion’s Eye Diamond reprint did (that card fell from 183 tickets to 60 with a promo reprint of only 2000 copies, then down to 12 with VMA).

To get 2000 copies into circulation would require of the order of thirty million treasure chest openings. MTGO is bigger than people think, but it is not that big.

Expect a price impact that is more like the effect that Zendikar Expeditions had on the existing cards on MTGO – ultra-low supply cards like Kor Haven and Dust Bowl did fall, but even low supply cards like Horizon Canopy and Cascade Bluffs were not impacted much.


Expected Value and Variance of Treasure Chests

Most analysis of Chests has been predicated on the assumption that the chests are worth almost nothing. This is wrong.

Expected Value is a term used to describe the average results of a random event, ignoring variance. For example, roll 3 dice and add the numbers showing, and your expected value is 10.5 (even though you cannot get exactly 10.5 at all, and the probability of getting exactly 10 is 12.5%, with exactly 11 also only being 12.5%).

For example, when Jace, Vryn’s Prodigy was 94 tickets, he added 0.75 tickets to the EV of opening an ORI pack, even though over 99% of ORI packs did not contain a copy of the card.

Variance can be ennumerated via a statistical term called the Standard Deviation, but I won’t go into that here other than to state that variance on Chests is very high.

The EV of a chest under my above assumptions is given by:

  • 0 value for the chaff (Standard commons/uncommons). It’s not technically 0 (you might get a Reflector Mage) but it’s close enough to be just a rounding error.
  • 1.25 ‘good items’ – that’s 0.625 Modern rares/mythics; 0.4375 bundles of 10 PP, and 0.1875 curated cards.
  • Each Modern rare or mythic has an average value of about 0.7 tickets (thanks to the guy on Reddit that determined that 0.8 tickets dealer sell price is the average at present – his post is here). This may seem hard to believe as a large majority of Modern legal rares are less than five cents. But the EV is propped up by the large number of 5+ ticket cards in the format and in particular by the 20+ ticket cards you will rarely see.
  • So the Modern rare/mythic slot contributes about 43c to the EV of a chest.
  • The PP slot contributes 44c under our assumptions.
  • Finally the curated cards can’t be estimated because we do not know their relative rarities. The mean value of a curated card ignoring their rarity weighting is about 5 tickets (thanks to BenBuzz790 on Reddit for this information), so I will estimate 2.5 tickets once we factor in rarity weighting (which may be miles off but it’s the best estimate we can do for now). That puts the curated card value also at around 45c.

So under these assumptions the EV for an opened chest is actually $1.35. The assumptions could be wrong, but I don’t see these errors tremendously changing the overall EV.

Once more the variance is high. Not as high as entering an Origins draft was back when Jace was 94 tickets, but it is still high, and it will not be rare to open 8 chests and to find the best things you got were 20 PP and an Anger of the Gods.


EV of events

Under the assumption that chests are worth 1.3 tickets, the ‘friendly’ events have the following absolute EVs at various win percentages:

50%: -10.6 PP

55%: -1.4 PP

60%: +8.15 PP

65%: +18 PP

70%: +28.4 PP


And under the further assumption that boosters are worth 10 tickets per draft set (an assumption made because of the draft entry fee changes occurring at the same time), the competitive events have the following absolute EVs:

50%: -13.5 PP

55%: +14.8 PP

60%: +46.3 PP

65%: +80 PP

Because the competitive events have better competition I will not give figures for 70%. That win rate is not sustainable.

Note that this provides a strong incentive for good players to get out of the kiddie pool and to swim with the sharks. This is definitely a good thing for MTGO’s future as it helps people who aren’t yet at an elite level play weaker opponents while getting tournament experience.



Improving the Chests

The four things I think Wizards should do with the Chests to make them more appealing are:

  • Make the damn things tradeable. Some people find a thrill in variance, others hate it. Let people make their own decision on what level of variance they will tolerate. Otherwise you will get a flood of rants from people that win 16 chests and open pure shit in them, and these people will speak louder than the people that open 3 chests and get a Port and a Liliana.
  • Publish transparent rarity information on everything in them. I helped Goatbots design the EV calculator after stopped being updated, and it is a tool that a lot of people use. Magic players like to know the odds and are usually suspicious of any situation in which they do not know the odds.
  • Consider replacing some of the PP ‘consolation prizes’ in casual leagues with fair amounts of Chests. I think everyone that enters a friendly league should get a consolation prize of a chest instead of the 10PP consolation prize you get for a 0-5 record.
  • Reduce some of the complexity in explaining what you get in them. I don’t see any need to have the 1-in-200 possibility of triple ‘prize’ packs – this should just be rolled into a better chance of getting a double prize pack.

Anyways this is pushing 2000 words now so I’ll leave it there and will respond to Reddit comments and comments on this site over time.


  • sirgog


5,081 total views, 2 views today

3 thoughts on “MTGO’s ‘Treasure Chests’ and competitive online play”

  1. You really should put some effort on making this site look decent, i like these articles but the layout and color scheme of the site makes me want to throw up.

    1. Heh. I agree – I’m clueless at web design and am spending a very small amount of time messing around with it.

      It’s probably going to continue to look shit in the medium term.

Leave a Reply

Your email address will not be published. Required fields are marked *